Many Internet services employ a pool of servers to process user requests. A client may initially be authenticated by a first server, but a future request from the client may end up on a different server that has not authenticated the client. As a result, the different server may re-authenticate the client by requesting authentication credentials once again from the user or the server may route the request to the first server that originally authenticated the client.